Fused lasso with a non-convex sparsity inducing penalty

Ilker Bayram, Po Yu Chen, Ivan Selesnick

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The fused lasso problem involves the minimization of the sum of a quadratic, a TV term and an ℓ1 term. The solution can be obtained by applying a TV denoising filter followed by soft-thresholding. However, soft-thresholding introduces a certain bias to the non-zero coefficients. In order to prevent this bias, we propose to replace the ℓ1 penalty with a non-convex penalty. We show that the solution can similarly be obtained by applying a modified thresholding function to the result of the TV-denoising filter.

Original languageEnglish (US)
Title of host publication2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages4156-4160
Number of pages5
ISBN (Print)9781479928927
DOIs
StatePublished - 2014
Event2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014 - Florence, Italy
Duration: May 4 2014May 9 2014

Other

Other2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014
CountryItaly
CityFlorence
Period5/4/145/9/14

Keywords

  • audio denoising
  • Fused lasso
  • non-convex penalty
  • thresholding
  • total variation denoising

ASJC Scopus subject areas

  • Signal Processing
  • Software
  • Electrical and Electronic Engineering

Cite this

Bayram, I., Chen, P. Y., & Selesnick, I. (2014). Fused lasso with a non-convex sparsity inducing penalty. In 2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014 (pp. 4156-4160). [6854384] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2014.6854384

Fused lasso with a non-convex sparsity inducing penalty. / Bayram, Ilker; Chen, Po Yu; Selesnick, Ivan.

2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 4156-4160 6854384.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Bayram, I, Chen, PY & Selesnick, I 2014, Fused lasso with a non-convex sparsity inducing penalty. in 2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014., 6854384, Institute of Electrical and Electronics Engineers Inc., pp. 4156-4160, 2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014, Florence, Italy, 5/4/14. https://doi.org/10.1109/ICASSP.2014.6854384
Bayram I, Chen PY, Selesnick I. Fused lasso with a non-convex sparsity inducing penalty. In 2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 4156-4160. 6854384 https://doi.org/10.1109/ICASSP.2014.6854384
Bayram, Ilker ; Chen, Po Yu ; Selesnick, Ivan. / Fused lasso with a non-convex sparsity inducing penalty. 2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 4156-4160
@inproceedings{9977a232a5704dd1ad2f37a702a9fe60,
title = "Fused lasso with a non-convex sparsity inducing penalty",
abstract = "The fused lasso problem involves the minimization of the sum of a quadratic, a TV term and an ℓ1 term. The solution can be obtained by applying a TV denoising filter followed by soft-thresholding. However, soft-thresholding introduces a certain bias to the non-zero coefficients. In order to prevent this bias, we propose to replace the ℓ1 penalty with a non-convex penalty. We show that the solution can similarly be obtained by applying a modified thresholding function to the result of the TV-denoising filter.",
keywords = "audio denoising, Fused lasso, non-convex penalty, thresholding, total variation denoising",
author = "Ilker Bayram and Chen, {Po Yu} and Ivan Selesnick",
year = "2014",
doi = "10.1109/ICASSP.2014.6854384",
language = "English (US)",
isbn = "9781479928927",
pages = "4156--4160",
booktitle = "2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Fused lasso with a non-convex sparsity inducing penalty

AU - Bayram, Ilker

AU - Chen, Po Yu

AU - Selesnick, Ivan

PY - 2014

Y1 - 2014

N2 - The fused lasso problem involves the minimization of the sum of a quadratic, a TV term and an ℓ1 term. The solution can be obtained by applying a TV denoising filter followed by soft-thresholding. However, soft-thresholding introduces a certain bias to the non-zero coefficients. In order to prevent this bias, we propose to replace the ℓ1 penalty with a non-convex penalty. We show that the solution can similarly be obtained by applying a modified thresholding function to the result of the TV-denoising filter.

AB - The fused lasso problem involves the minimization of the sum of a quadratic, a TV term and an ℓ1 term. The solution can be obtained by applying a TV denoising filter followed by soft-thresholding. However, soft-thresholding introduces a certain bias to the non-zero coefficients. In order to prevent this bias, we propose to replace the ℓ1 penalty with a non-convex penalty. We show that the solution can similarly be obtained by applying a modified thresholding function to the result of the TV-denoising filter.

KW - audio denoising

KW - Fused lasso

KW - non-convex penalty

KW - thresholding

KW - total variation denoising

UR - http://www.scopus.com/inward/record.url?scp=84905259007&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84905259007&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2014.6854384

DO - 10.1109/ICASSP.2014.6854384

M3 - Conference contribution

SN - 9781479928927

SP - 4156

EP - 4160

BT - 2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -